PAPIv2 Key [disk] not supported by GCP backend about workflows

[a-frantz]

Thanks for filing an issue, but I'm not sure what you're asking for.

Our disk key specifies the minimum storage size for a task running on a DNAnexus/Azure VM. It's expected for other backends to ignore this value (keys in the runtime section of a WDL task can be backend specific, so being ignored by alternate backends is expected. For example, we often run our pipeline on our local LSF cluster, which similarly logs a message about ignoring the disk key).

I was unable to find GCP documentation about the disks key you are referring to, but I'm guessing it means something other than what disk on DNAnexus/Azure means.

Are the workflows failing, or just logging this message? If they're failing, what is the specific error causing the failure?

[TomGardner]

Thanks for the quick response. Ok, given the WARNING it is probably ignored and won't cause issues.
I put in my options file:

    "default_runtime_attributes": {
        "zones": "us-central1-a us-central1-b us-central1-c us-central1-f",
        "memory": "16G",
        "disks": "local-disk 100 SSD"
    }

But I see in the logs:
Adjusting boot disk size to 12 GB: 10 GB (runtime attributes) + 1 GB (user command image) + 1 GB
Which indicates to me either the 'disk' is ignored or my 'disks' is ignored.

[TomGardner]

Perhaps this is a Cromwell issue.

[a-frantz]

Sorry, I don't think anyone on this team has experience with GCP. This sounds like a Cromwell configuration issue on your end that I don't think we're equipped to handle.

However the below documentation links seem extensive, and may help you:
https://cromwell.readthedocs.io/en/stable/tutorials/PipelinesApi101/
https://cromwell.readthedocs.io/en/stable/backends/Google/

[TomGardner]

Ok. I can close. Thanks for looking at it.